First measurements of reactive α-dicarbonyl concentrations on PM 2 . 5 aerosol over the Boreal forest in Finland during HUMPPA-COPEC 2010 – source apportionment and links to aerosol aging

The first dataset for summertime boreal forest concentrations of two atmospherically relevant α-dicarbonyl compounds, glyoxal (Gly) and methylglyoxal (Mgly) on PM2.5 aerosol was obtained during the HUMPPA-COPEC2010 field measurement intensive in Hyyti älä, Finland. Anthropogenic influences over the course of the campaign were identified using trace gas signatures and aerosol particle chemical composition analysis. The data evaluation allowed the identification of different events such as urban pollution plumes, biomass burning and sawmill emissions as sources of high Gly and Mgly concentrations. Mean aerosol concentrations during periods of biogenic influence were 0.81 ng m−3 for Gly and 0.31 ng m−3 for Mgly. Mgly was generally less abundant in PM 2.5, probably due to its shorter photolysis lifetime and less effective partitioning into the particle phase due to its smaller effective Henry’s Law constant compared to Gly. This is in contrast with previous urban studies which show significantly more Mgly than Gly. Peak concentrations for Gly coincided with nearby sources, e.g. high VOC emissions from nearby sawmills, urban pollution plumes from the city of Tampere located 50 km southwest of the sampling site and biomass burning emissions from wildfires. Calculated ratios of Gly in PM 2.5 and total organic matter in PM1 aerosols indicate higher values in less aged aerosols. Irreversible processing of Gly in the particle phase, e.g. via oxidation by OH radicals, organo sulfate or imidazole formation are processes currently discussed in the literature which could likely explain these findings.


Introduction
Atmospheric aerosols are a major source of uncertainty in estimations of the global radiative forcing (IPCC, 2007). Improving knowledge about sources, formation, growth and transformation processes of secondary organic aerosols (SOA) is necessary to reduce these uncertainties, as SOA makes up a large fraction of submicron atmo-global secondary organic aerosol (SOA) budget (Kulmala et al., 2004b;Tunved et al., 2006), since it represents over one quarter of the world's forest area (FAO, 2010). In contrast to tropical forests, where isoprene is the most abundant BVOC in the boundary layer (e.g. Lelieveld et al., 2008), monoterpenes (e.g. α-and β-pinene, ∆ 3 -carene), especially α-pinene, are often the most abundant gas phase compounds emitted by the Boreal forest (e.g. Ebben et al., 2011;Hakola et al., 2003).
During summer 2010 the HUMPPA-COPEC 2010 field measurement intensive was conducted from 12 July to 12 August at the SMEAR II measurement site in Hyytiälä (Hari and Kulmala, 2005), Finland. Key features of the station, land use near the site, predominant wind directions and meteorological conditions are described in de- 15 tail by Williams et al. (2011). Briefly, the SMEAR II research station is located at 61 • 50 50.685 N and 24 • 17 41.206 E at 179 m above sea level (a.s.l.). The land within a radius of 50 km of the measurement site is dominated by coniferous and mixed forests with only one significant urban center (Tampere) to the southwest. During the campaign air masses originated primarily from southwestern (53.7 %) and southeastern Introduction urban pollution plume events, with regard to glyoxal (Gly) and methylglyoxal (Mgly) concentrations in particulate matter with aerodynamic diameters <2.5 µm (PM 2.5 ), will also be investigated. The two simplest α-dicarbonyl compounds have drawn much attention as potential SOA precursors over the last years. They are produced by a variety of biogenic 5 (e.g. isoprene or monoterpenes) and anthropogenic (e.g. benzene, toluene, xylenes) sources (Fu et al., 2008). Even though their vapor pressures are too high to produce SOA with regard to traditional gas/particle partitioning theory (Odum et al., 1996;Seinfeld and Pankow, 2003) they are readily taken up by aqueous aerosols and water containing particles due to the high solubility of their mono-and di-hydrates in water.
The Gly/Mgly measurements presented here analyze Gly/Mgly concentrations in PM 2.5 particles as the sum of the monomeric α-dicarbonyl, its hydrated forms and re-25 versibly formed higher molecular weight compounds, using a derivatization technique as will be discussed in the measurements section. Thus, not the complete SOA formed via aqueous phase chemistry of Gly and Mgly in atmospheric particulates is measured. However, the compounds analyzed here specifically account for Gly/Mgly SOA, while some of the aforementioned compounds, e.g. oxalic or formic acid and other oxidation products, can have multiple precursors. The purpose of this paper is the identification of non-natural Gly/Mgly influences and the quantification of background forest Gly/Mgly concentrations using aerosol and trace gas measurements during HUMPPA-COPEC 2010 in combination with backward 5 trajectories. The aging state of ambient aerosols will be assessed through identification of specific aerosol sources and Gly/Mgly concentrations. Therefore, analysis of Gly and Mgly aerosol concentrations in combination with aerosol chemical composition analysis via FTIR, and trace gas measurements of carbon monoxide (CO), sulfur dioxide (SO 2 ), acetonitrile (CH 3 CN), n-pentane and important BVOC's (α-, β-pinene, ∆ 3 -carene, isoprene) will be used for the appointment of sources and investigations about the aging state of the respective aerosols. Additionally, biogenic pristine forest background concentrations of Gly and Mgly in Boreal aerosols are quantified. 15 A parallel set of two sampling devices was operated one platform below the top of the radiation tower (16 m a.g.l.) at the Boreal forest research station SMEAR II (Station for Measuring Ecosystem-Atmosphere Relation) in Hyytiälä, Finland, from 10 July to 10 August 2010. Each sampling device consisted of a size selective inlet (Digitel DPM PM 2.5 , Riemer Messtechnik, Germany), which was operated at a flow rate of 2. HYSPLIT model) for the wind vectors northwest (NW), northeast (NE), southeast (SE) and southwest (SW). The second sampling device (B) was used as a backup and for additional experiments. Sampling durations were 8 to 57 h. Loaded filters were stored in capped glass vessels covered with aluminum foil at −20 • C until analysis. HPLC-ESI-MS/MS analysis of glyoxal and methylglyoxal as their bis-2,4-dinitrophenylhydrazone 5 derivatives was carried out using an Agilent 1100 series HPLC system (degasser, gradient pump, auto sampler, diode array detector, Agilent Technologies GmbH, Germany) equipped with a Varian Pursuit XRs 3 C8 (Varian, Germany) reversed phase analytical column and a Bruker HCT-Plus ion trap mass spectrometer (Bruker Daltonics, Germany) as described by Kampf et al. (2011). Briefly, loaded filters were extracted by means of ultra-sonication, concentrated under a gentle N 2 -stream and derivatized using a freshly cleaned acidic aqueous solution of 2,4-dinitrophenylhydrazine (DNPH). Reversibly formed higher molecular weight compounds are efficiently de-oligomerized due to the strong acidic conditions used. Hydration of α-dicarbonyl monomers is a fast equilibrium reaction, thus constantly providing non hydrated carbonyl functional 15 groups needed for the derivatization reaction. Therefore, under these conditions the α-dicarbonyl compounds, their hydrated forms and reversibly formed oligomers yield one product upon derivatization, i.e. glyoxal-bis-2,4-dinitrophenylhydrazone (Gly-bis-DNPH) and methylglyoxal-bis-2,4-dinitrophenylhydrazone (Mgly-bis-DNPH), respectively. Specific fragments of the protonated ions were quantified for each target com-20 pound in multiple reaction monitoring mode (MRM) of the mass spectrometer.

Measurements of anthropogenic trace gases
Pentane measurements were obtained with the Fast Observation of Trace Organics System (FOTOS). The FOTOS consists of a custom, three-trap (water, enrichment, and cryofocusing) cryogenic sampling system, a fast GC (20 m × 0.25 mm × 1.0 µm, 25 PLOT-Al 2 O 3 "S"), and a MS quadrupole detector (Agilent 5973 The SO 2 and CO concentrations were measured from different levels (4.2, 8.4, 16.8, 33.6, 50.4 and 67.2 m) in the 73 m high tower of the SMEAR II site. For a description of the site and a more detailed description of the instrumentation see Hari and Kulmala (2005) and Kulmala et al. (2001). SO 2 concentration was measured with a UV-fluorescence analyzer (TEI 43 CTL, Thermo Fisher Scientific, Waltham, MA, 5 USA) in sequence from all the six levels. Time resolution was 1 min at 6 min intervals for each measurement level. Detection limit was 0.1 ppb and relative accuracy ±3 %. CO mixing ratios were measured on top of the HUMPPA tower (24 m) using a fast-response vacuum-UV resonance fluorescence instrument (Model AL-5002, Aero Laser, Germany) (Gerbig et al., 1999). The instrument was calibrated in the field every 10 three hours with a secondary standard (260 ± 2.1 ppbv CO in compressed air) traced to a NOAA reference standard. The time resolution of the original data is 1 s. Detection limit and total uncertainty were 1 ppbv and <10 %, respectively.

GC-MSD analysis of BVOC concentrations (pinenes, carene, isoprene)
BVOCs were observed with an on-line sampling and TD-GC-MSD measurement sys-  . Detection limits were below 10 pptv. More details about the instrument description are given in Williams et al. (2007).

FTIR sample collection and chemical composition analysis
Submicron particles were collected 4 m above ground, approximately mid-canopy level, in temperature controlled housing on 37 mm Teflon filters (Pall Inc., 37 mm diameter, 5 1.0 µmm pore size) located downstream of a 1 µmm sharp-cut cyclone (SCC 2.229 PM1, BGI Inc.). The aerosol was dried to <75 % RH (Silica orange gel 13767, Sigma Aldrich, USA) prior to filter collection. Back filters were also collected to quantify VOC absorption and artifacts; however, absorption on all back filters was below limit of detection (LOD). Filter duration varied from 6, 9, and 12 h, including duplicate 24-h filters, 10 which were used when mass loadings were low. To reduce evaporative losses, samples were stored and transported at 0 • C prior to analysis in San Diego, California, by a Tensor 27 spectrometer (Bruker, Billerica, MA). An automated peak-fitting algorithm described by Russell et al. (2009) was used to quantify organic functional groups, which included saturated aliphatic C-CH (alkane), unsaturated aliphatic C=CH (alkene), aro-15 matic C=CH, non-acid organic hydroxyl C-OH (alcohol), primary amine C-NH 2 , nonacid carbonyl C=O (aldehyde and ketone), and carboxylic acid COOH groups. Alkene and aromatic groups were below the limit of quantification (LOQ) for the entirety of the field campaign.
3 Results and discussion 20

Overview of α-dicarbonyl concentrations during HUMPPA-COPEC 2010
Field data concentrations of Gly and Mgly in PM 2.5 during HUMPPA-COPEC 2010 are provided from 10 July to 10 August. The concentrations ranged from 0.54 to 12.07 ng m −3 for Gly and from below limit of detection (LOD) to 1.26 ng m −3 for Mgly. Variation of Mgly concentration was much less pronounced as illustrated in Fig. 1 sampling according to the wind vectors as described in the previous section allowed the data to be analyzed for the potential sources of high Gly and Mgly concentrations. High concentrations of Gly and Mgly were observed from southwestern (SW) and southeastern (SE) wind vectors, while northwestern (NW) wind vectors did not show elevated concentrations. Land usage characteristics suggest a stronger anthro-5 pogenic influence in the south of the sampling site (Williams et al., 2011), whereas the NW sector was deemed to be most representative of the boreal forest ecosystem. Several distinct sources for high α-dicarbonyl concentrations are discussed in detail in the following section.
The mean concentration of Gly over the entire campaign was calculated to be 10 1.78 ± 2.14 ng m −3 . If the peak concentrations are considered to be of anthropogenic origin, as will be shown in the following section, and left aside, the mean concentration attributed to the natural biogenic emissions of the surrounding coniferous forest was 0.81 ± 0.23 ng m biogenic emissions. However, Mgly concentrations are surprisingly low considering that most likely major precursors for Gly and Mgly in this area are monoterpenes and isoprene. Both are abundant and possess a relatively high molar Mgly formation yield upon their atmospheric oxidation (e.g. 25 and 4.2 % molar yield upon isoprene and monoterpene oxidation for Mgly, compared to 6.2 and 2.8 % for Gly) (Fu et al., 2008). 25 However, modeled gas phase summertime mixing ratios of Gly and Mgly for southern Finland in 2006 appear to be similar, 10-20 ppt for Gly and 5-20 ppt for Mgly (Fu et al., 2008). This is likely due to the shorter lifetime of Mgly with respect to photolysis and OH radical oxidation. Henry's Law constants (K H ), describing the flux of solutes between gas and aqueous phases (Betterton et al., 1992), and a shorter chemical lifetime may also contribute to the explanation of lower Mgly aerosol concentrations. Effective Henry's Law constants including hydration of the carbonyl groups (K * H ) reported for Gly and Mgly at room temperature are 3.6 × 10 5 and 3.7 × 10 3 M atm −1 (Zhou and Mopper, 1990;Betterton and Hoffmann, 1988), respectively, indicating uptake into the 5 aerosol phase to be more important for Gly.
Other field data from boreal forest environments are very limited. Liggio et al. (2003) report Gly and Mgly PM concentrations in Vancouver and other rural and biogenic influenced sites (Sumas mountain) in Canada, ranging from 0.43 to 3.3 ng m −3 in the case of Gly and from 0.03 to 1.2 ng m −3 in the case of Mgly. These results are in good agreement with the concentrations encountered during HUMPPA-COPEC 2010. Gly peak concentrations are higher in the recent field campaign, owing to the diverse anthropogenic events occurring over the course of the campaign, as will be discussed in the following sections. Field data from major cities are reported by Kawamura and Yasui (2005), Li and Yu (2005) and Ho et al. (2007). The samples were collected in 15 Hong Kong, other major Chinese cities and Tokyo, Japan. Gly and Mgly PM concentrations reported are one to two orders of magnitude higher than reported from boreal forest regions. Maximum concentrations of Gly and Mgly in these cities were up to 72.9 and 218.4 ng m −3 , respectively. Interestingly, the reported urban Mgly PM concentrations are high compared to Gly PM concentrations in major cities, while in the boreal 20 forest environments Gly concentrations appear to be higher and more variable than Mgly concentrations. Mgly was also more abundant in a small field study conducted in southwest Germany at the Taunus Observatory, which is influenced by continental emission particularly from the highly industrialized Rhein-Main area (Kampf et al., 2011). Therefore, high concentrations of Mgly may suggest a stronger anthropogenic 25 influence. However, isoprene-emitting forests (e.g. Amazonian rain forests, Guenther et al., 2006) might show higher values of Mgly compared to Gly due to a higher molar yield of Mgly from atmospheric isoprene oxidation processes (Fu et al., 2008). This should be addressed in future field studies.

Clean air, pollution events, sawmill activity and wildfires
Throughout the campaign several biomass burning, sawmill influence and urban/anthropogenic pollution events have been identified by tracer gases (Williams et al., 2011) and via chemical composition of the aerosol phase from FTIR. Biomass burning events, i.e. the Russian wildfires which began from mid July 2010, were identified by sharp increases in the acetonitrile (CH 3 CN) and carbon monoxide (CO) mixing ratios, as well as high non-acid carbonyl and hydroxyl functional group concentrations in the aerosol phase. Saw mill influences were detected as peaks in BVOC concentrations accompanied by minima in the (-)/(+)-α-pinene ratio and urban/anthropogenic influences were identified as peaks in sulfur dioxide (SO 2 ) and n-pentane time traces. Figure 2 illustrates the influences of these events during the collection of the filter samples for Gly/Mgly analysis. On several occasions high Gly and Mgly concentrations coincide with the aforementioned events. On 12 July, an event with high sub-micron particle number counts, which was deduced to be a pollution plume from the city of Tampere in the southwest 15 of the sampling site (Williams et al., 2011), both Gly and Mgly concentrations show a local maximum. Ensemble 4 day back trajectories of the events attributed to have an urban/ anthropogenic origin on 14 July (St. Petersburg, Russia), 16 July (Poland and Germany), 26 July to 30 July (partly anthropogenic influence from St. Petersburg, Russia coincident with Russian wildfire emissions) and 6 August (unclear origin, but 20 identified as anthropogenic via peak in n-pentane mixing ratio) are available in the supporting material (Figs. S1 to S5 in the Supplement). These events barely show any elevated Mgly concentration, which might be due to the greater distance of the respective sources from the sampling site. More distant sources result in longer transport times, thus increasing the chemical processing of α-dicarbonyls to corresponding acids and other later generation reaction products in the aerosol phase. This reduces the observable concentrations close to the background level and will be examined again in the following Sect. 3.3. measurement intensive was measured a few hours after the sawmill event on 6 August. The sawmill event on 18 July on the other hand is not visible with regard to increased Gly concentrations. Air masses, however, originated mainly from NW wind vectors in 5 this case, in contrast to SW wind vectors during the events on 3 and 6 August, thus originating from a different source. The biomass burning event from 7 to 9 August resulted in increased Gly concentrations of up to 8.23 ng m −3 , while no such increase is visible in the period from 26 to 30 July. This indicates that a relatively fresh plume was directly transported to the 10 sampling site from 7 to 9 August. Ensemble 4 day back trajectories and fire hotspot data for the August event are available in the supporting material (see Figs. S8-S10). On 7 August the transport time from the wildfires in the Russian Federation to the sampling site was around 3 to 4 days coinciding with the lowest Gly concentration for this event. On 8 August a wildfire with a transport time <24 h appears in the back trajectory 15 plot, which potentially caused the high Gly concentrations observed on that day. The transport time from wildfires to sampling site on 9 August was around 2 days. Other parameters affecting the aging process of the aerosol potentially changed between the two periods of wildfire influencing the sampling site.
High Gly concentrations are also visible at the beginning of the campaign on 10 July. 20 Unfortunately no data for the specific marker compounds are available for this time. However, we attribute these high concentrations to another pollution plume from the city of Tampere to the southwest of the sampling site, since the air mass backward trajectories calculated using the NOAA HYSPLIT model (see Fig. S7) indicate that the advected air masses went over Tampere. An increased vehicular activity at the 25 measurement site itself associated with setting up instruments potentially may also contribute to the high Gly concentrations observed. Acetylene, which forms Gly and not Mgly and was presumably emitted by vehicles at the site, was identified as a potential SOA precursor by Volkamer et al. (2009). The pollution source seems to be different Introduction

Tables Figures
Back Close

Full Screen / Esc
Printer-friendly Version

Interactive Discussion
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | two days later (12 July), as on this day also an increase in Mgly concentration was observed. Nucleation events are usually observed during spring and fall at the SMEAR II station, though some occasionally occur also during the summer season when air masses originate over the Arctic Ocean (Kulmala et al., 2004a;Sogacheva et al., 2005). On 23 5 July a nucleation event was observed during the HUMPPA-COPEC campaign (Williams et al., 2011). Interestingly, Gly concentrations show a local minimum on this day. This finding can potentially be attributed to the very clean air with a campaign minimum of total submicron particle counts (3-1000 nm) at 06:00 LT.

Comparison with FTIR data
10 FTIR analysis was used to characterize the chemical composition of PM 1 aerosols during HUMPPA-COPEC 2010. A comparison of organic matter (OM) characterized as hydroxyl (-OH) functional groups and α-dicarbonyl concentrations could offer useful information. It should be noted that Gly and Mgly are present in the aerosol phase as their hydrated forms (mono-or di-hydrate) or oligomers formed via hemiacetal, acetal, 15 imine, anhydride, ester, and sulfate ester formation and aldol condensation (see Lim et al., 2010 and references therein), thus drastically reducing free carbonyl functional groups related to Gly or Mgly and increasing hydroxyl functional groups at the same time. Concentrations of hydroxyl functional groups in PM 1 and Gly concentrations in PM 2.5 are illustrated in Fig. 3. 20 During the biomass burning events from 26 to 30 July a significant amount of hydroxyl functional groups is present in the aerosol phase with a peak concentration of 3.05 µg m −3 on 29 July. Only minor changes in Gly PM concentrations are visible during that time, suggesting that most of the reactive α-dicarbonyl compounds have already been irreversibly processed. During the biomass burning event on 8 August high con-  Table 1. Maximum Gly/OM ratios were observed during the sawmill events on 3 and 6 August. In these less aged aerosols Gly makes up 0.2 and 0.41 % of the total OM respectively. Biomass Burning events on 26 and 29 July were addressed as strongly aged aerosols in the previous discussion; this is also reflected here by low Gly/OM ratios of 0.01 %. On 10 8 August on the other hand the values are slightly higher, likely due to direct transport of a fresh plume during this event. At selected days without anthropogenic influences and wind vectors from SW (21 July) and NW (23 and 25 July) directions the Gly/OM ratio ranged from 0.03 to 0.14 %. The high value was observed on the day of the nucleation event, thus again suggesting higher values in the Gly/OM ratio are related to less aged 15 aerosols. Gly/-OH ratios show similar trends with respect to potential aerosol aging state. On 6 August Gly/-OH reached a campaign maximum of 1.78 %, also the other sawmill event on 3 August showed a high value of 1.43 %, while 0.2-0.4 % for the clean air days and 0.03-0.04 % for the first two and 0.16 % for the third biomass burning event were observed, respectively.

Time resolution and diurnal behavior of Gly in PM 2.5
In general filter sampling for Gly and Mgly was conducted related to air mass origins. While one of the sampling devices was used that way throughout the whole campaign, the other one was used to investigate the diurnal concentration behavior of Gly and Mgly in the week from 15 to 19 July. However, as mentioned before Mgly concentrations 25 were very low resulting in total Mgly masses per filter below the LOD during this week of experiments. Gly concentrations in PM 2.5 for the week from 15 to 19 July are illustrated in Fig. 4 A strong diurnal variation of the Gly PM concentrations is visible, however, without a clear trend. On 15 and 16 July the night time samples show higher concentrations than the day time samples, while the opposite is true for 17 and 18 July. Wind sectors changed quite frequently during that time (see Fig. 4). The analysis of air mass origins showed high day and low night time concentrations for Gly from the NW wind sec-5 tor, while the concentration behavior for Gly in air masses from southern wind sectors seems to be more complex. Comparisons between this temporal behavior and that of tracer measurements like aliphatic hydrocarbons (n-pentane), sulfur dioxide (SO 2 ), carbon monoxide (CO) and nitrogen oxides (NO x ) as anthropogenic tracers and BVOC's (α-, β-pinene, isoprene, ∆ 3 -carene and myrcene) as biogenic tracers, however, did not 10 show significant correlation.
The Gly measurements with short and long sampling durations were found to be consistent with each other. Each ambient sample with a long sampling duration shows the mean concentration of the two corresponding ambient samples with short sampling durations with a maximum error of <6 %, thus demonstrating the reliability of the an- 15 alytical method applied and suggesting sampling artifacts to be minor. The drawback of a longer sampling duration, however, is obvious. Diurnal cycles in the concentration behavior are not observed due to longer averaging and short spikes or minima in concentration become invisible.

20
HUMPPA-COPEC 2010 offered the opportunity to acquire the first set of data for PM 2.5 concentrations of the two most prominent α-dicarbonyl compounds glyoxal and methylglyoxal in the boreal forest region of Finland. These concentrations were compared to a variety of other field data in order to appoint sources and investigate the relative abundance and relevance of α-dicarbonyl compounds in the Boreal environment. Mean 25 background concentrations of Gly and Mgly at the SMEAR II station in Hyytiälä, Finland, were 0.81 and 0.31 ng m −3 , respectively. However, local and regional events, e.g. biomass burning, sawmill emissions and urban/anthropogenic pollution, showed strong influences on α-dicarbonyl aerosol concentrations and resulted in peak Gly concentrations of up to 12.07 ng m −3 . From a mass perspective Gly and Mgly seem not as relevant in boreal forest environments as in highly anthropogenic influenced urban centers (compare Volkamer et al., 2007). At least 15 % of the SOA formation in Mex- 5 ico City was attributed to glyoxal, which corresponds to several µg m −3 . Mgly aerosol phase concentrations in the boreal forest are low in comparison to previous urban measurements. This may be due to the short photolysis lifetime and less pronounced uptake by SOA particles and a higher abundance of aromatic precursors with greater Mgly yields, e.g. xylenes, in urban environments. Comparison with FTIR PM 1 chemical composition data allowed the aging state of aerosols advected from distinct events and sources to be assessed qualitatively by Gly/OM and Gly/-OH ratios. While short distanced events and events with great source strengths had a strong influence and high Gly/OM ratios up to 0.4 %, more distant and smaller events were not visible via increased α-dicarbonyl concentrations and thus showed low Gly/OM ratios of 0.01 %. 15 On background days, Gly/OM ranged from 0.03-0.14 %. Low Gly concentrations in aged aerosols are reasonable since Gly is known to irreversibly form imidazoles and other high molecular weight compounds in ammonium containing aerosols (Galloway et al., 2009;Noziere et al., 2009;Shapiro et al., 2009) and to be processed into glyoxylic acid and further to oxalic acid, formic acid and CO 2 through OH radical reactions 20 and as side products in imidazole formation in the aqueous phase (Carlton et al., 2007;Lim et al., 2010;Yu et al., 2011).

Tables Figures
Back Close

Full Screen / Esc
Printer-friendly Version

Interactive Discussion
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | is acknowledged for providing the fire data used in this publication. C. J. K. and T. H. acknowledge the Max Planck Institute for Chemistry in Mainz for financial support and the possibility to participate in the HUMPPA-COPEC-2010 field measurement intensive. A. L. C. and L. M. R. acknowledge support from NSF grant ATM-0904203, and ALC was supported in part by the Department of Energy Office of Science Graduate Fellowship Program (DOE SCGF) administered 5 by ORISE-ORAU under contract no. DE-AC05-06OR23100. The campaign measurements and analyses were supported by the European Integrated Project on Aerosol Cloud Climate and Air Quality Interactions EUCAARI (project No. 036833-2), the EUSAAR TNA (project No. 400586). TP acknowledges funding from the Academy of Finland (1118615, 139656) and from Maj and Tor Nessling foundation (2010143,2012443).